{"url":"/task/passage-retrieval","name":"Passage Retrieval","slug":"passage-retrieval","description_markdown":"**Passage retrieval** is a specialized type of IR application that retrieves relevant passages (or pieces of text) rather than an entire ranked set of documents.","categories":[{"name":"Natural Language Processing","url":"/area/natural-language-processing"}],"source":{"archive":"pwc-archive (Hugging Face), CC BY-SA 4.0","snapshot":"2025-07-28","slug_source":"archive_url"},"counts":{"papers_tagged":268,"papers_with_code":133,"benchmarks":5,"benchmark_tables_in_archive":19,"benchmark_tables_shown":19,"benchmark_tables_withheld_as_spam":0,"benchmark_definition":"a leaderboard table with at least one row; benchmark_tables_shown also counts the zero-row tables; benchmark_tables_in_archive adds the tables withheld as spam","datasets":10,"subtasks":0,"parent_tasks":1},"benchmarks":[{"leaderboard":"/sota/passage-retrieval-on-msmarco-beir","slug":"passage-retrieval-on-msmarco-beir","dataset":"MSMARCO (BEIR)","dataset_url":"/dataset/beir","rows_in_archive":12,"metrics":["nDCG@10"],"first_row_in_archive_order":{"model":"BM25+CE","paper_title":"BEIR: A Heterogenous Benchmark for Zero-shot Evaluation of Information Retrieval Models","paper_url":"/paper/beir-a-heterogenous-benchmark-for-zero-shot","paper_date":"2021-04-17","arxiv_id":"2104.08663","code_links":[{"title":"osu-nlp-group/hipporag","url":"https://github.com/osu-nlp-group/hipporag"},{"title":"UKPLab/beir","url":"https://github.com/UKPLab/beir"},{"title":"beir-cellar/beir","url":"https://github.com/beir-cellar/beir"}],"syntology":{"n":4,"n_ran":3,"n_unverified":1,"n_pointer_only":0}}},{"leaderboard":"/sota/passage-retrieval-on-natural-questions","slug":"passage-retrieval-on-natural-questions","dataset":"Natural Questions","dataset_url":"/dataset/natural-questions","rows_in_archive":10,"metrics":["Precision@100","Precision@20"],"first_row_in_archive_order":{"model":"ReAtt","paper_title":"Retrieval as Attention: End-to-end Learning of Retrieval and Reading within a Single Transformer","paper_url":"/paper/retrieval-as-attention-end-to-end-learning-of","paper_date":"2022-12-05","arxiv_id":"2212.02027","code_links":[{"title":"jzbjyb/reatt","url":"https://github.com/jzbjyb/reatt"}],"syntology":null}},{"leaderboard":"/sota/passage-retrieval-on-peerqa","slug":"passage-retrieval-on-peerqa","dataset":"PeerQA","dataset_url":"/dataset/peerqa","rows_in_archive":8,"metrics":["Recall@10","MRR"],"first_row_in_archive_order":{"model":"SPLADEv3","paper_title":"SPLADE-v3: New baselines for SPLADE","paper_url":"/paper/splade-v3-new-baselines-for-splade","paper_date":"2024-03-11","arxiv_id":"2403.06789","code_links":[{"title":"naver/splade","url":"https://github.com/naver/splade"}],"syntology":{"n":3,"n_ran":3,"n_unverified":0,"n_pointer_only":3}}},{"leaderboard":"/sota/passage-retrieval-on-entityquestions","slug":"passage-retrieval-on-entityquestions","dataset":"EntityQuestions","dataset_url":"/dataset/entityquestions","rows_in_archive":7,"metrics":["Recall@20"],"first_row_in_archive_order":{"model":"TOME-2","paper_title":"Mention Memory: incorporating textual knowledge into Transformers through entity mention attention","paper_url":"/paper/mention-memory-incorporating-textual-1","paper_date":"2021-10-12","arxiv_id":"2110.06176","code_links":[{"title":"google-research/language","url":"https://github.com/google-research/language/tree/master/language/mentionmemory"}],"syntology":null}},{"leaderboard":"/sota/passage-retrieval-on-ms-marco-1","slug":"passage-retrieval-on-ms-marco-1","dataset":"MS MARCO","dataset_url":"/dataset/ms-marco","rows_in_archive":1,"metrics":["MAP"],"first_row_in_archive_order":{"model":"TW-BERT","paper_title":"End-to-End Query Term Weighting","paper_url":"/paper/end-to-end-query-term-weighting","paper_date":"2023-08-06","arxiv_id":null,"code_links":[{"title":"ledmaster/tw-bert","url":"https://github.com/ledmaster/tw-bert"}],"syntology":null}},{"leaderboard":null,"slug":"passage-retrieval-on-mmarco-ar","dataset":"mMARCO-ar","dataset_url":null,"rows_in_archive":0,"metrics":["MRR@10","Recall@10","Recall@100","Recall@1000","Recall@500"],"first_row_in_archive_order":null},{"leaderboard":null,"slug":"passage-retrieval-on-mmarco-de","dataset":"mMARCO-de","dataset_url":null,"rows_in_archive":0,"metrics":["MRR@10","Recall@10","Recall@100","Recall@1000","Recall@500"],"first_row_in_archive_order":null},{"leaderboard":null,"slug":"passage-retrieval-on-mmarco-en","dataset":"mMARCO-en","dataset_url":null,"rows_in_archive":0,"metrics":["MRR@10","Recall@10","Recall@100","Recall@1000","Recall@500"],"first_row_in_archive_order":null},{"leaderboard":null,"slug":"passage-retrieval-on-mmarco-es","dataset":"mMARCO-es","dataset_url":null,"rows_in_archive":0,"metrics":["MRR@10","Recall@10","Recall@100","Recall@1000","Recall@500"],"first_row_in_archive_order":null},{"leaderboard":null,"slug":"passage-retrieval-on-mmarco-fr","dataset":"mMARCO-fr","dataset_url":null,"rows_in_archive":0,"metrics":["MAP@10","MRR@10","Recall@10","Recall@100","Recall@1000","Recall@500","nDCG@10"],"first_row_in_archive_order":null},{"leaderboard":null,"slug":"passage-retrieval-on-mmarco-hi","dataset":"mMARCO-hi","dataset_url":null,"rows_in_archive":0,"metrics":["MRR@10","Recall@10","Recall@100","Recall@1000","Recall@500"],"first_row_in_archive_order":null},{"leaderboard":null,"slug":"passage-retrieval-on-mmarco-id","dataset":"mMARCO-id","dataset_url":null,"rows_in_archive":0,"metrics":["MRR@10","Recall@10","Recall@100","Recall@1000","Recall@500"],"first_row_in_archive_order":null},{"leaderboard":null,"slug":"passage-retrieval-on-mmarco-it","dataset":"mMARCO-it","dataset_url":null,"rows_in_archive":0,"metrics":["MRR@10","Recall@10","Recall@100","Recall@1000","Recall@500"],"first_row_in_archive_order":null},{"leaderboard":null,"slug":"passage-retrieval-on-mmarco-ja","dataset":"mMARCO-ja","dataset_url":null,"rows_in_archive":0,"metrics":["MRR@10","Recall@10","Recall@100","Recall@1000","Recall@500"],"first_row_in_archive_order":null},{"leaderboard":null,"slug":"passage-retrieval-on-mmarco-nl","dataset":"mMARCO-nl","dataset_url":null,"rows_in_archive":0,"metrics":["MRR@10","Recall@10","Recall@100","Recall@1000","Recall@500"],"first_row_in_archive_order":null},{"leaderboard":null,"slug":"passage-retrieval-on-mmarco-pt","dataset":"mMARCO-pt","dataset_url":null,"rows_in_archive":0,"metrics":["MRR@10","Recall@10","Recall@100","Recall@1000","Recall@500"],"first_row_in_archive_order":null},{"leaderboard":null,"slug":"passage-retrieval-on-mmarco-ru","dataset":"mMARCO-ru","dataset_url":null,"rows_in_archive":0,"metrics":["MRR@10","Recall@10","Recall@100","Recall@1000","Recall@500"],"first_row_in_archive_order":null},{"leaderboard":null,"slug":"passage-retrieval-on-mmarco-vi","dataset":"mMARCO-vi","dataset_url":null,"rows_in_archive":0,"metrics":["MRR@10","Recall@10","Recall@100","Recall@1000","Recall@500"],"first_row_in_archive_order":null},{"leaderboard":null,"slug":"passage-retrieval-on-mmarco-zh","dataset":"mMARCO-zh","dataset_url":null,"rows_in_archive":0,"metrics":["MRR@10","Recall@10","Recall@100","Recall@1000","Recall@500"],"first_row_in_archive_order":null}],"datasets":[{"url":"/dataset/natural-questions","name":"Natural Questions","full_name":"","num_papers_in_archive":1404},{"url":"/dataset/ms-marco","name":"MS MARCO","full_name":"Microsoft Machine Reading Comprehension Dataset","num_papers_in_archive":1036},{"url":"/dataset/beir","name":"BEIR","full_name":"Benchmarking IR","num_papers_in_archive":311},{"url":"/dataset/entityquestions","name":"EntityQuestions","full_name":"","num_papers_in_archive":30},{"url":"/dataset/qampari","name":"QAMPARI","full_name":"","num_papers_in_archive":19},{"url":"/dataset/peerqa","name":"PeerQA","full_name":"","num_papers_in_archive":12},{"url":"/dataset/germandpr","name":"GermanDPR","full_name":"","num_papers_in_archive":2},{"url":"/dataset/coresearch","name":"CoreSearch","full_name":"","num_papers_in_archive":1},{"url":"/dataset/csprd","name":"CSPRD","full_name":"Chinese Stock Policy Retrieval Dataset","num_papers_in_archive":1},{"url":"/dataset/dapfam","name":"DAPFAM","full_name":"A Domain‑Aware Patent Retrieval Dataset Aggregated at the Family Level","num_papers_in_archive":1}],"subtasks":[],"parent_tasks":[{"url":"/task/information-retrieval","name":"Information Retrieval"}],"papers":{"order":"repositories listed in the archive (desc), then date (desc); the archive holds no stars","population":"papers tagged with this task that list at least one repository in the archive","shown":30,"of":133,"tagged_in_all":268,"items":[{"url":"/paper/dense-passage-retrieval-for-open-domain","title":"Dense Passage Retrieval for Open-Domain Question Answering","date":"2020-04-10","arxiv_id":"2004.04906","repositories_listed":19,"syntology":{"n":14,"n_ran":10,"n_unverified":4,"n_pointer_only":9}},{"url":"/paper/leveraging-passage-retrieval-with-generative","title":"Leveraging Passage Retrieval with Generative Models for Open Domain Question Answering","date":"2020-07-02","arxiv_id":"2007.01282","repositories_listed":8,"syntology":null},{"url":"/paper/towards-unsupervised-dense-information","title":"Unsupervised Dense Information Retrieval with Contrastive Learning","date":"2021-12-16","arxiv_id":"2112.09118","repositories_listed":6,"syntology":{"n":5,"n_ran":1,"n_unverified":4,"n_pointer_only":0}},{"url":"/paper/passage-re-ranking-with-bert","title":"Passage Re-ranking with BERT","date":"2019-01-13","arxiv_id":"1901.04085","repositories_listed":6,"syntology":{"n":13,"n_ran":2,"n_unverified":11,"n_pointer_only":0}},{"url":"/paper/approximate-nearest-neighbor-negative","title":"Approximate Nearest Neighbor Negative Contrastive Learning for Dense Text Retrieval","date":"2020-07-01","arxiv_id":"2007.00808","repositories_listed":5,"syntology":{"n":12,"n_ran":3,"n_unverified":9,"n_pointer_only":0}},{"url":"/paper/bm25s-orders-of-magnitude-faster-lexical","title":"BM25S: Orders of magnitude faster lexical search via eager sparse scoring","date":"2024-07-04","arxiv_id":"2407.03618","repositories_listed":3,"syntology":{"n":22,"n_ran":10,"n_unverified":12,"n_pointer_only":0}},{"url":"/paper/drop-your-decoder-pre-training-with-bag-of","title":"Drop your Decoder: Pre-training with Bag-of-Word Prediction for Dense Passage Retrieval","date":"2024-01-20","arxiv_id":"2401.11248","repositories_listed":3,"syntology":{"n":2,"n_ran":2,"n_unverified":0,"n_pointer_only":0}},{"url":"/paper/dapr-a-benchmark-on-document-aware-passage","title":"DAPR: A Benchmark on Document-Aware Passage Retrieval","date":"2023-05-23","arxiv_id":"2305.13915","repositories_listed":3,"syntology":null},{"url":"/paper/zusammenqa-data-augmentation-with-specialized","title":"ZusammenQA: Data Augmentation with Specialized Models for Cross-lingual Open-retrieval Question Answering System","date":"2022-05-30","arxiv_id":"2205.14981","repositories_listed":3,"syntology":null},{"url":"/paper/colbertv2-effective-and-efficient-retrieval","title":"ColBERTv2: Effective and Efficient Retrieval via Lightweight Late Interaction","date":"2021-12-02","arxiv_id":"2112.01488","repositories_listed":3,"syntology":{"n":1,"n_ran":1,"n_unverified":0,"n_pointer_only":0}},{"url":"/paper/beir-a-heterogenous-benchmark-for-zero-shot","title":"BEIR: A Heterogenous Benchmark for Zero-shot Evaluation of Information Retrieval Models","date":"2021-04-17","arxiv_id":"2104.08663","repositories_listed":3,"syntology":{"n":4,"n_ran":3,"n_unverified":1,"n_pointer_only":0}},{"url":"/paper/never-lost-in-the-middle-improving-large","title":"Never Lost in the Middle: Mastering Long-Context Question Answering with Position-Agnostic Decompositional Training","date":"2023-11-15","arxiv_id":"2311.09198","repositories_listed":2,"syntology":{"n":17,"n_ran":15,"n_unverified":2,"n_pointer_only":0}},{"url":"/paper/fine-tuning-llama-for-multi-stage-text","title":"Fine-Tuning LLaMA for Multi-Stage Text Retrieval","date":"2023-10-12","arxiv_id":"2310.08319","repositories_listed":2,"syntology":{"n":9,"n_ran":8,"n_unverified":1,"n_pointer_only":0}},{"url":"/paper/knowledge-distillation-of-large-language","title":"MiniLLM: Knowledge Distillation of Large Language Models","date":"2023-06-14","arxiv_id":"2306.08543","repositories_listed":2,"syntology":null},{"url":"/paper/i-3-retriever-incorporating-implicit","title":"I^3 Retriever: Incorporating Implicit Interaction in Pre-trained Language Models for Passage Retrieval","date":"2023-06-04","arxiv_id":"2306.02371","repositories_listed":2,"syntology":{"n":10,"n_ran":0,"n_unverified":10,"n_pointer_only":0}},{"url":"/paper/expand-rerank-and-retrieve-query-reranking","title":"Expand, Rerank, and Retrieve: Query Reranking for Open-Domain Question Answering","date":"2023-05-26","arxiv_id":"2305.17080","repositories_listed":2,"syntology":null},{"url":"/paper/query-as-context-pre-training-for-dense","title":"Query-as-context Pre-training for Dense Passage Retrieval","date":"2022-12-19","arxiv_id":"2212.09598","repositories_listed":2,"syntology":null},{"url":"/paper/dptdr-deep-prompt-tuning-for-dense-passage","title":"DPTDR: Deep Prompt Tuning for Dense Passage Retrieval","date":"2022-08-24","arxiv_id":"2208.11503","repositories_listed":2,"syntology":null},{"url":"/paper/contextual-mask-auto-encoder-for-dense","title":"ConTextual Masked Auto-Encoder for Dense Passage Retrieval","date":"2022-08-16","arxiv_id":"2208.07670","repositories_listed":2,"syntology":null},{"url":"/paper/qampari-an-open-domain-question-answering","title":"QAMPARI: An Open-domain Question Answering Benchmark for Questions with Many Answers from Multiple Paragraphs","date":"2022-05-25","arxiv_id":"2205.12665","repositories_listed":2,"syntology":{"n":1,"n_ran":1,"n_unverified":0,"n_pointer_only":0}},{"url":"/paper/dureader-retrieval-a-large-scale-chinese","title":"DuReader_retrieval: A Large-scale Chinese Benchmark for Passage Retrieval from Web Search Engine","date":"2022-03-19","arxiv_id":"2203.10232","repositories_listed":2,"syntology":null},{"url":"/paper/multi-cpr-a-multi-domain-chinese-dataset-for","title":"Multi-CPR: A Multi Domain Chinese Dataset for Passage Retrieval","date":"2022-03-07","arxiv_id":"2203.03367","repositories_listed":2,"syntology":null},{"url":"/paper/large-dual-encoders-are-generalizable","title":"Large Dual Encoders Are Generalizable Retrievers","date":"2021-12-15","arxiv_id":"2112.07899","repositories_listed":2,"syntology":null},{"url":"/paper/salient-phrase-aware-dense-retrieval-can-a","title":"Salient Phrase Aware Dense Retrieval: Can a Dense Retriever Imitate a Sparse One?","date":"2021-10-13","arxiv_id":"2110.06918","repositories_listed":2,"syntology":null},{"url":"/paper/slot-filling-for-biomedical-information","title":"Slot Filling for Biomedical Information Extraction","date":"2021-09-17","arxiv_id":"2109.08564","repositories_listed":2,"syntology":null},{"url":"/paper/robust-retrieval-augmented-generation-for","title":"Robust Retrieval Augmented Generation for Zero-shot Slot Filling","date":"2021-08-31","arxiv_id":"2108.13934","repositories_listed":2,"syntology":{"n":8,"n_ran":1,"n_unverified":7,"n_pointer_only":0}},{"url":"/paper/dealing-with-typos-for-bert-based-passage","title":"Dealing with Typos for BERT-based Passage Retrieval and Ranking","date":"2021-08-27","arxiv_id":"2108.12139","repositories_listed":2,"syntology":null},{"url":"/paper/learning-to-retrieve-how-to-train-a-dense","title":"Learning To Retrieve: How to Train a Dense Retrieval Model Effectively and Efficiently","date":"2020-10-20","arxiv_id":"2010.10469","repositories_listed":2,"syntology":null},{"url":"/paper/open-domain-question-answering-goes","title":"Open-Domain Question Answering Goes Conversational via Question Rewriting","date":"2020-10-10","arxiv_id":"2010.04898","repositories_listed":2,"syntology":{"n":8,"n_ran":0,"n_unverified":8,"n_pointer_only":0}},{"url":"/paper/overview-of-the-trec-2019-deep-learning-track","title":"Overview of the TREC 2019 deep learning track","date":"2020-03-17","arxiv_id":"2003.07820","repositories_listed":2,"syntology":null}],"syntology_records":14,"syntology_note":"a paper without a record is not a recorded non-run: it may lack an arXiv id or simply be absent from the graph layer"},"description_links":{"kept":0,"unwrapped_to_text":0,"bare_urls_linked":0,"relative_images_dropped":0,"rule":"internal links are kept only when the target slug exists in the catalog"},"syntology":{"read_at":"2026-09-24T18:15:14+00:00","claim":"Per-sample execution status on synthesized fixtures ('ran N of M samples'); not a correctness claim and not a ranking signal.","status_vocabulary":{"ran_honours":"ran, honoured the contract we drafted","ran_violates":"ran, violated the contract we drafted","ran_draft_wrong":"ran; our contract draft was wrong, not the code","ran_fixture":"ran; our fixture could not drive it","ran":"ran on a synthesized input","unverified":"unverified (harvested, no recorded run)"}},"not_shown":{"libraries":"the archive has no per-task library table","trend_sparklines":"the Trend column of the benchmarks table was a rendered image; it is not in the archive","social_and_latest_sorts":"stars and social signals are not in the archive"}}